Papers tendProceedings of the Royal Society of Tasmania, Volume 133(3), 2000 51

BOTTOM WATER FORMATION AND POLYNYAS IN ADEllE LAND, ANTARCTICA

by Nathaniel L. Bindoff, Stephen R. Rintoul and Robert Massom

(with five text-figures)

BINDOFF, NL., RINTOUL, S.R. & MASSOM, R., 2000 (31 :v): Bottom water formation and polynyas in Adelie Land, Antarctica. In Banks, M.R. & Brown, M.J. (Eds): TASMANIAAND THE SOUTHERN OCEAN. Pap. Proc. R. Soc. T asm. 133(3): 51-56. ISSN 0080- 470). Antarctic CRC, GPO Box 252-80, Hobart, Tasmania, Australia 7001 (NLB, SRR, RM); and CSIRO Division of Marine Research, GPO Box 1538, Hobart, Tasmania, Australia 7001 (SRR).

Antarctic B~ttom Water is the coldest and densest water found in the global ocean. It spreads into all the major ocean basins, carrying the cold water towards the equatorial regions, and is a central component of the global thermo-haline circulation. However, the mechanisms of bottom water formation are not well established; its geographical distribution and rate of formation have yet to be fully quantified. Polynyas, which are large persistent openings in sea-ice that form during the winter near the Antarctic Coast, playa central role in the formation or Antarctic Bottom Water. This paper describes the bottom water formation around the Antarctic continental margin with particular emphasis on the processes and mechanisms of the Adelie Land Bottom Water formation near Dumont D'Urville south of Tasmania. Key Words: Antarctic Bottom Water, polynya, brine rejection, Adelie Land Bottom Water.

INTRODUCTION Australian-Antarctic basin than in either the Weddell or ~ ~. Ross Seas. Antarctic Bottom Water (AABW) is one of the most Bottom waters are an important part of the thermo­ important water-masses in the global ocean. It is the coldest, haline circulation, and increased greenhouse gas scenarios densest water found in the deep ocean. This water is found from coupled-ocean-atmosphere models show that the rate in all of the deep basins around the Antarctic continent, of bottom water formation is expected to decrease (Manabe broken only by the relatively shallow sill through Drake et al. 1990, Manabe et al. 1991). It is important that these Passage. Estimates of the volume of AABW (defined by model scenarios are validated with observations. In this water denser than neutral density - Jackett & McDougall paper the geographical distribution of ADLBW in the 1997 [yn> 28.27 kg.m-3]) show it to occupy 3.5% of the Australian-Antarctic basin, evidence for its time variability volume of the ocean (Orsi et al. 1999) and affect, through of formation and its links to the Mertz Polynya are described. mixing and advection, more than 41 % of the global oceanic volume (Worthington 1981). Estimates of the production rate of bottom water vary but range from 8-12 Sv (Orsi et BOTTOM WATERS IN THE al. 1999 [1 Sv=106 m3.s-1]), giving a renewal time (defined AUSTRALIAN-ANTARCTIC BASIN here as the volume occupied by AABW divided by production rate) of 180-120 years, respectively. This renewal time is Describing the precise distribution of bottom waters and quite long, reflecting the fact that the production rate is estimating their production are essential requirements for small compared with the volume of the ocean occupied by understanding the mechanisms and sources ofbottom water AABW. Yet it is this production rate combined with the formation. The CFC-ll concentration at 10 m above the relatively low mixing rates which creates the strong density ocean floor (fig. 1) is taken from hydrographic data obtained gradients that define the Antarctic Circumpolar Current. on a voyage of the RSV Aurora Australis from January­ The main formation of AABW is thought to occur in March 1996 (see Bindoff et al. 1997 and Rosenberg et al. three distinct regions, on the continental shelves of the 1997 for a more complete description of this voyage). These western Weddell Sea, western Ross Sea and Adelie Land, CFC-ll measurements are accurate to ~ 1% (WOCE south of Tasmania. The Adelie Land source of bottom standard). Although the source ofCFC-ll in the atmosphere water has always been considered to be negligible (Carmack is increasing, the rate of increase from 1990 to 1996 is just 1977) compared with the Weddell and Ross Sea sources. 6% or about 1% per year (Gras et aL 1999). Thus, for the However, recent analyses of the volume of the AABW relatively young waters directly above the ocean floor and characteristic of these three source regions show that the the shelf waters presented here « 10 years old) there is no Weddell, Adelie and Ross sources are respectively 68%, need to adjust the CFC-ll values to allow for the time 24% and 8% by volume (Rintoul 1998), that is Adelie variation in the atmospheric source, and for these data this Land Bottom Water (ADLBW) is the second largest source. adjustment is small compared to the effects of mixing. The This new result is also supported by a census of Chloro­ CFC-ll Chloro-fluorocarbons are transferred from the fluorocarbon (CFC-ll) concentrations around the Antarctic atmosphere to the oceans through the surface of the ocean. continent (Orsi et al. 1999), which shows that layer Because CFCs are chemically passive in the ocean, they act of AABW defined by the CFC-ll is thicker in the as a tracer or dye. The highest concentrations (to first order) 52 NL. Bindoff ,S,R. Rintoul and R. Massom

Bottom CFC11 (pm/kg) 62

63

64

66

67

6~0 90 100 110 120 130 140 150 160 Longitude °E

FIG. 1 -~ The CFC11 concentration (pmollkg) taken from the deepest water sample (typically 10 m from bottom) for all eTD stations over the deep ocean. The numbers are the observed CFC-11 concentration at each CTD. The size ofthe circle gives the amplitude of'the CFC11 concentration. The 1000 and 3000 m depth contours are shown. These data come from the MARGINEX experiment (Bindoffet al. 1997), obtained in January-March 1996. The thick dashed contours show the ice shelves in this region.

can be interpreted here as waters that have been most dense) water flowing downslope in either of these two recently in contact with the atmosphere. Each of the north­ sections. This suggests that the source waters are flowing south hydrographic sections (except at 1500E) cross the down the continental slope during the winter or between continental shelfbreak shown by the 500 m isobath. On the the two sections or in discrete canyons on the continental shelf floor, very high CFC-ll concentrations of greater than slope (Rintoul 1998). 4 pmol.kg-1 occur, consistent with the moderately rapid The seasonal variability of bottom water formation is overturning of the shelf waters and mixing with the poorly known because the extensive sea-ice cover during atmosphere. However, these concentrations decrease very winter makes the continental slope and shelf region largely rapidly down the continental slope. In the deep ocean, the inaccessible to conventional ship-based measurements. lowest deep-water CFC-ll concentrations occur in the west However, temperature measurements from 10 m above the (along 800E) with concentrations less than 1 pmol.kg-1, and ocean bottom from a mooring at 65°S, 1400E in 2600 m these concentrations progressively increase eastward until a of water show a distinctive seasonal signal. The warmest local maximum in CFC-ll concentration occurs at 1400E. temperatures occur during the February-June period and At 1500E the CFC-1 I concentration decreases again. the strongest cooling in the August-December period These lower CFC-l1 concentrations are also accompanied (Fukamachi, pers. comm.); this is consistent with the by higher salinities and warmer temperatures, consistent strongest formation being during the late winter-spring. with this water originating from the Ross Sea and flowing In addition, the temperature and salinity characteristics westwards along the continental rise (Gordon & Tchernia of shelf waters during the January-March 1996 voyage of 1972). The increase in CFC-11 at 1400E also accompanies the RV Aurora Australis show that the shelf salinities (fig. 2A) water that is colder, fresher and more oxygen-rich, implying are too fresh to form bottom waters, supporting the that the bottom waters at 1400E have been mixed with conclusions from the moored temperature measurements. more recently ventilated water originating from the Although the Ross Sea Bottom Water (RSBW) has continental shelf somewhere between 140° and 1500E. distinctive temperature and salinity characteristics present Excluding the large values of CFC-11 over the continental at 1500E, there is no evidence for this signature at 128°E shelf! slope break (less than 1000 m), all of the north-south (fig. 2A), where the bottom waters are colder and fresher sections show the highest values of CFC-11 (and also the (labelled as ADLBW). Here, the temperature-salinity coldest, freshest and highest in oxygen) offshore in waters correlation for waters < O°C forms a straight line that is deeper than 3000 m (fig. 1). Although it appears from noise free (fig. 2A). The shelf waters have a temperature these data that the source must be between 140° and near the surface freezing temperature (-1.85°C) and a 1500E, it is clear that during the summer time there is not salinity less than 34.5 psu. For this section there is no a continuous plume of high CFC (cold and fresh and simple two-end member mixing scheme between the Bottom water formation and polyn,yas in Antarctica 53 summer shdf waters with the Modified Circumpolar Deep and 0.5°C) and is present along the entire slope between Water (labelled MCDW) and the ADLBW. By contrast, 140° and 1500 E (fig. 3). With the exception of the Adelie the CF C-temperature correlation shows a different Depression (142°-145°E), this temperature maximum is relationshi? between the shelf water andADLBW (fig. 2B). always colder than -l.O°C for hydrographic casts south of The she1fVlaters ar(-1.85°C), the MCDW and the bottom the 500 m isobath, indicating that very little of the warm waters (

Polynyas, which are large persistent openings in sea-ice that ON-SHELF TRANSPORT OF MODIFIED form during the winter near the Antarctic Coast, playa CIRCUMPOLAR DEEP WATER central role in the formation of AABW. They are active regions of sea-ice formation, because the cold Antarctic There are two processes for increasing the salinity of shelf winds «-30°C and average speeds in excess of 25 m.s-1) waters. The first is through the transport of relatively warm freeze the surface waters and then blow this newly formed ice and salty MCDW (fig. 2A) onto the continental shelf away from the coast. It is this transport of sea-ice from the (primarily through channels through the shelf break) and coast and the shape of the coastline that keep the polynyas the second is through brine rejection from the formation of open and make these open areas a factor of ten greater in ice sea-ICe. production compared with ocean covered in sea-ice (Zwally Although there are relatively few hydrographic observa­ eta!. 1985). tions on the continental shelf region between 140° and The high release of heat to the atmosphere and the high 150 0 E, there is some evidence from earlier measurements sea-ice production both act to increase the density of the that during summer, MCDW does penetrate onto the waters below the polynya. In some polynyas, where the continental shelf. MCDW is characterised by a shallow bottom topography has an ideal shape, the brine rejection temperature maximum (and salinity maximum). Offshore during ice-formation in combination with the on-sheil' over the continental slope (north of the 500 m isobath) the flow of saline waters from offshore increases the salinity of temperature maximum is relatively warm (between -0.5° the water column. When the shelf salinity reaches a critical

8r-----~----~------~ 8 o o

6 ,~OShelfWater 0> ~ : ~ %8- "'-E \8 : \ - 0. , 00 "'~'MCDW' .. ..- 4 " " .. " , ~ ':" " .,.... ~ o () :, 0 @:> LL - \ o ...... DLB () : \ 0 ci o (L -1 .. ,<;q , , , , , , ,-" "'"""", ,:, , , ooR ...... ' .... gj , 2 "'ADl~W: '\\ife'~Q'.0:0 : 0 0: cP - - o : -0 0 : " boo - - 0 -_-0 ClD o : 00 o~ n..

value, t~ese shelf waters become dense enough to flow to be the greatest ice production in East Antarctica (Cavalieri dowa t~e continental shelf to the abyss. The principal & Martin 1985). reaso:ns [or the uncertainties in bottom water formation are The Mertz Polynya lies immediately above the Adelie that .Jirect observations of the processes are difficult to Depression (cf. fig. 4 with fig. 1). This polynya forms to make from ship and satellite, and that the strongest the west of the Mertz Glacier. Its size is also partly controlled formatim period is believed to be when the sea-ice is at its by the line of grounded icebergs and fast ice that extends greatest extent around the Antarctic continent (i.e. north from this glacier (at 146°E), blocking the westward Septe mber-October). flow of sea-ice into the sea-ice area. The katabatic winds are Brine rejection during sea-ice formation also plays a role strongly offshore at the coastline, turning westwards in 20- in inerelsing the winter time salinity. A detailed census of 30 km. However, the ice-free region extends well beyond the significant polynyas between 40° and 1600 E over an the zone of strong winds, and this polynya is wider than eight-yetr period, using satellite passive microwave data, most (Massom et al. 1998). The unusual width of this shows that the largest and most persistent polynyas are the polynya is also associated with this region becoming ice­ Shackleton Ice Shelf and the Mertz Polynyas with respect­ free earlier in the spring than the regions immediately to ively an average area in winter of 30 000 and 23 000 km2 the east and west (Gloersen et al. 1992, Massom et al. (Massom et al. 1998). However, in spite of its larger size 1998) and may be the result of the upwelling of warm the Shackleton Ice Shelf (near 95°E) is not associated with MCDW in the polynya region. This polynya has a strong bottom water formation (fig. 1). The factors that affect seasonal signal (fig. 5) with its areal extent becoming largest production are not the size of the polynya alone, but also typically in October and ranging from 20 000 to include the strength and direction of the local winds, air 60 000 km2 for this month. The fastest rate of growth in temperature and water temperature. Along the Antarctic size of this polynya tends to occur during August­ coastline, the Mertz Polynya (67"5, 145°E) is in a region September, which is consistent with maximum sea-ice of some of the world's strongest and most persistent winds production and coincides with the colder bottom waters (Adolphs & Wendler 1995, Ball 1957). It is these factors observed in temperature time-series measured offshore that give rise to the Mertz Polynya having what is believed (Fuchamachi, pers. com.).

-65

-65.5

-66

-66.5

(I,) "'0 ...... :::l :p -67 CCI -67.5 • -0.5 < T

0 -1.0 < T < -0.5 -68 + -1.5 < T < -1 .0 -68.5 .. T < -1.5 -69 140 145 150 155 160 longitude FIG. 3 - The potential temperature of the potential temperature maximum ofMCD W plotted using the available historical data. The: 500, 1000, 2000 and 3000 m isobaths are shown. Note that the warmest MCDW on the she/fis found in the Adilie depressim1 between 143° and 145°E. Bottom water formation and po/ynyas in Antarctica 55

CONCLUSIONS relatively more saline) MCDW waters, thereby enhancing the overturning circulation within the Adelie Depression From th esedata, the Mertz Polynya appears to playa central caused by the outflow ofcold bottom water. In this conceptual role in t:heformation of ADLBW. It provides additional model, the Mertz Polynya is being kept open by the winds salinity t:hrough high sea-ice production and, because of the (i.e. acting as a latent heat polynya) and also by heat cold air te1l1peratures, provides a window for high heat­ transported onshore by MCDW (i.e. acting as a sensible fluxes from the ocean to the atmosphere. These high heat­ heat polynya). The fact that no MCDW is observed on the fluxes can cool the likely onshore transport of warm (and continental shelf except in the Adelie Depression also

FIG. 4 --An AVHRR channel 4 (thermal infrared) image ofthe Antarctic coastline taken in August 1995 between 150° and 135°E, south of Tasmania. The Mertz Po/ynya is the dark region to the west ofthe Mertz Glacier and adjacent to the Antarctic coastline. The grey-scale gives an indication of the temperature, with the brightest greys corresponding to the coldest reflective surfoce (glacier ice or clouds) and the darkest regions to open ocean.

60 N ., 75% ice cone. E .. 65% ice cone. ~ (Y) 40 - .. (;) or- ...... X 20 ('Q i! « 0 r I r :< t i JASO JJAS 0 JJ ASO J J A S 0 1987 1988 1989 1991 Months FIG. 5 - Eight-year time series ofthe area ofthe Mertz Po/ynya as determined from the Special Sensor Microwave Image satellite passive microwave observations. The po/ynya is deemed to be an open area when the sea-ice concentration falls below a specified concentration. In this figure, the open area was determined for three cases where the concentrations are respective/y less than 75%, 65% and 55%. From Massom et al. (1998) 56 HI. Bindoff, S.R. Rintoul and R. Massom indicates that the shape of the Adelie Depression and its GORDON, A.L. & TCHERNIA, P., 1972: Waters of the continental associated channel through the shelf break is a critical margin offAdelie coast, Antarctica. In Hayes, D.E. (Ed.): element in the formation of ADLBW. ANTARCTIC OCEANOLOGYII: THEAUSTRALIAN­ The essential questions that need to be resolved are the NEW ZEALAND SECTOR. Antarct. Res. Ser. American precise role that the onshore transport of MCDW plays Geophysical Union, Washington: 59-69. GRAS, ].L., DEREK, N., TINDALE, N.W. & DICK, A.L., 1999: compared with the role of ice production within the Mertz BASELINE ATMOSPHERIC PROGRAM AUSTRALIA Polynya and the rate of production of Addie Land Bottom 1996. Bureau of Meteorology and CSIRO Division of Water. These questions are being addressed through a Atmospheric Research: 146 pp. multidisciplinary programme involving sea-ice scientists, JACKETT, D.R. & McDOUGALL, T.J., 1997: A neutral density meteorologists and oceanographers in July 1999. variable for the world's ocean. J Phys. Oceanogr. 27(2): 237-263. MANABE, S., BRYAN, K. & SPELMAN, M.J., 1990: Transient response ACKNOWLEDGEMENTS of a global ocean-atmosphere model to a doubling of atmospheric carbon dioxide. J Phys. Oceanogr. 20: 722- The hard work of the Master and crew on the RV Aurora 749. MANABE, S., STOUFFER, R.J., SPELMAN, M.J. & BRYAN, K., 1991: Australis is greatly appreciated. The CFC data were kindly Transient responses of a coupled ocean-atmosphere model made available by Dr Mark Warner, School ofOceanography, to gradual changes of atmospheric CO2 • Part I : annual University of Washington, U.S.A. This work is a contri­ mean response. J Climate 4: 785-818. bution to the W orid Ocean Circulation Experiment and the MASSOM, R., HARRIS, P., MICHAEL, K. & POTTER, M., 1998: The CSIRO Climate Change Research Program. distribution and formative processes of latent heat polynyas in East Antarctica. Ann. Glac. 27: 420-426. ORSI, A.H. , JOHNSON, G. e. & BULLISTER, J.L., 1999: Circulation, REFERENCES mixing and production ofAntarctic Bottom Water. Progr. Oceanogr.: in press. RINTOUL, S.R., 1998: On the origin and influence of Adelie Land ADOLPHS, U. & WENDLER, G., 1995: A pilot study on the Bottom Water. In Jacobs, S. & Weiss, R. (Eds): OCEAN, interactions between katabatic winds and polynyas at the ICE,ANDATMOSPHERE:INTERACTIONSATTHE Adelie Coast, Eastern Antarctica. Antarct. Sci. 7: 307- 314. ANTARCTIC CONTINENTAL MARGIN. Antarct. Res. BALL, F.K., 1957: The katabatic winds of Adelie Land and King Ser. 75. American Geophysical Union, Washington: 151- 17l. George V Land. Tellus 9: 201-208. ROSENBERG, M., BRAY, S., BINDOFF, N.L., RINTOUL, S.R., BINDOFF, N.L., WARNER, M.J. & NICOL, S., 1997: The Antarctic & Margin Experiment. Int. WOCE News!. 26: 36-38. JOHNSTON, N., BELL, S. TOWLER, P., 1997: Aurora Australis marine science cruises AU9501, AU9604 and CARMACK, E.e., 1977: Water characteristics of the Southern Ocean south of the Polar Front. In Angel, M. (Ed.): AU9601 - oceanographic field measurements and VOYAGE OF DISCOVERY: GEORGE DEACON 70TH analysis, intercruise comparisons and data quality notes. Antarctic CRC, Hobart 12: 150 pp. ANNIVERSARY VOLUME. Deep-Sea Research, Pergamon Press, Oxford: 15-4l. WORTHINGTON, L.V., 1981: The water masses of the worlds ocean: some results of a fine-scale census. In Warren, CAVALIERI, D.J. & MARTIN, S., 1985: A passive microwave study B.A. & Wunsch, e. (Eds): EVOLUTION OF PHYSICAL of polynyas along the Antarctic Wilkes Land coast. In Jacobs, S.S. (Ed.): OCEANOLOGY OF THE OCEANOGRAPHY. MIT Press, Cambridge: 42-69. CONTINENTAL ANTARCTIC SHELF. Antarct. Res. ZWALLY, H.J., COMISO, J.e. & GORDON, A.L., 1985: Antarctic offshore leads and polynyas and oceanographic effects. Ser. American Geophysical Union, Washington: 227- 252. In Jacobs, S.S. (Ed.): OCEANOLOGY OF THE GLOERSEN, P., CAMPBELL, W.S., CAVALIERI, D.J., COMISO, J.e., ANTARCTIC CONTINENTAL SHELF. Antarct. Res. Ser. American Geophysical Union, Washington: 203- PARKINSON, e.L. & ZWALLY, H.J., 1992: Satellite passive microwave observations and analysis. National 226. Aeronautics and Space Administration, Washington, DC, (accepted 5 July 1999) NASA SP-511: 290 pp.